TWI552722B - R-wave detection algorithm using enhanced so and chan method - Google Patents

Description

R-wave detection algorithm using enhanced So and Chan method

The invention relates to a related field of electrocardiogram detection, in particular to an R wave detection algorithm using a reinforced So and Chan method.

ECG research has been more than 50 years old, and the algorithms for R-wave detection are numerous. There are many wavelet transform algorithms, Tompkins algorithms and dynamic threshold algorithms. The technology uses the So and Chan algorithm based on the dynamic threshold algorithm and is improved.

The So and Chan algorithm calculates the slope of each point of the acquired time domain data, then finds the maximum slope and sets the threshold to detect the QRS complex, and finally detects the position of the R wave, due to this algorithm. The R wave is detected by calculating the slope, so the noise of the baseline drift has little effect on it. The Modified So and Chan algorithm improves the RG inversion of the subject's ECG signal by squared the slope. If the square of the slope of two consecutive points is greater than the square of the slope threshold, the QRS_onest is determined. Effectively detects normal R wave and R wave inversion.

Please refer to FIG. 1 for a flow chart showing a conventional So and Chan algorithm. The So and Chan algorithm was mentioned in a paper published by H.H.So and K.L.Chan in 1997. This algorithm was developed to improve the paper published by Friesen et al. in 1990. The steps of the conventional So and Chan algorithm include: step SA1: providing electrocardiogram data; step SA2: calculating a slope according to the provided data; step SA3: obtaining a slope maximum value; step SA4: calculating a slope threshold; Step SA5: confirm whether the slope of two consecutive points is greater than the slope threshold; if not, return to step SA2; if yes, proceed to the next step; step SA6: detect the QRS complex start point (QRS onset); step SA7: The R wave maximum value (R-peak) is retrieved; and step SA8: the slope threshold is updated, and the process returns to step SA4.

Wherein, in the above step SA2, the So and Chan algorithm calculates the slope of each point of the acquired time domain data, as shown in the formula (1), slope ( n )=-2 X ( n -2)- X ( n -1) + X ( n +1) + 2 X ( n + 2) (1).

Then, in the above steps SA3 and SA4, the maximum slope is determined and the threshold is set to detect the QRS complex, and finally the position of the R wave is detected, because the algorithm detects the slope by calculating the slope. Waves, so the noise of the baseline drift has little effect on it. Calculate the slope threshold (slope_threshold) as shown in equation (2).

The threshold parameter (threshold_parameter) can be adjusted to 2, 4, 8, 16 according to the voltage amplitude of the electrocardiogram; the threshold value represents the ratio of the magnification of the maximum slope max i in the ith second signal, and the smaller the parameter, the pair of R waves It has higher sensitivity when detecting, but has poorer ability to resist noise. The larger the parameter, the lower the sensitivity to R wave, but the stronger the ability to resist noise. In this case, select 8 when the slope of two consecutive points When it is greater than the slope threshold, it represents the starting point of the QRS wave (QRS onset), and the maximum value can be found below to locate the R wave. Each time the R wave is detected, maxi is updated once. Equation (3) and formula (4) are as follows: First _maxi= Height of R po int- height of QRS onset (4).

The filter parameter (filter_param) can also be set to 2, 4, 8, and 16. Here, 16 is selected, and the initial value of maxi is preset to the maximum slope value in the first second signal.

This algorithm is applicable to the ECG monitoring system with mobility and immediacy, because it has two major features: 1. The algorithm has less computational complexity and is easier to implement on a wafer that can be detected immediately; 2. It has a pair of R Wave amplitude and adaptive filter parameters can increase its sensitivity.

However, if the noise is sharp and steep, it is easy to cause misjudgment. If the ECG waveform has R wave inversion (QS complex) and RSR' wave and T wave high tip, the detection ability is poor.

The Modified So and Chan algorithm was published by the Taipei University of Science and Technology in 2008. It was developed for the original So and Chan and added some techniques in the Tompkins algorithm. This method improves the RG inversion of the ECG signal of the subject. . The slope is calculated as a square. If the square of the slope of two consecutive points is larger than the square of the slope threshold, the QRS_onest is determined. This method can effectively detect the normal R wave and R wave inversion.

Please refer to FIG. 2, which is a flow chart showing a conventional Modified So and Chan algorithm. The steps of the Modified So and Chan algorithm include: step SB1: providing electrocardiogram data; step SB2: calculating slope; step SB3: obtaining slope maximum; step SB4: calculating slope threshold; step SB5: confirming whether the slope of two consecutive points is Greater than the slope threshold; if not, proceed to the next step; if yes, set the QRS complex start parameter (QRS onset) to True and set the maximum parameter (Max) to True (step SB51), then jump to Step SB6: Step SB6: confirm whether the square value of the slope of two consecutive points is greater than the square value of the slope threshold; if not, return to step SB2; if yes, set the QRS complex start parameter (QRS onset) to True And setting the maximum value parameter (Max) to False (step SB61), and proceeding to the next step; step SB7: confirming whether QRS onset is True; if yes, increasing the number of R wave waits (Wait_R) by 1 (step SB71), and Proceed to the next step; if not, proceed directly to the next step; step SB8: confirm whether Wait_R is equal to 100; if not, return to step SB2; if yes, proceed to the next step; step SB9: confirm the maximum parameter (Max) is No; True; if yes, the maximum value max_point is retrieved (step SB91), and the next step is performed; if not, the minimum value min_point is retrieved (step SB92), and the next step is performed; Step SB10: Set Wait_R to 0; and Step SB11: Update the slope threshold and end the flow.

Based on So and Chan and Modified So and Chan, a new method is proposed, that is, when an R wave is detected, the QRS_onset is paused for a period of time, and the R wave detection is not used after detecting the QRS_onset. The maximum value of the peak in a period of time, but the selection of the first turning point, this way can effectively improve the occurrence of false detection of RSR' wave and T wave too high; and the technology of the present invention reduces the Modified So and Chan The calculation of the square of the slope has been calculated. The absolute value of the slope is taken as an absolute value. This method can accurately detect the reverse R wave. When the reverse R wave is detected, the valve is not used. The value is updated.

The present invention provides an R wave detection algorithm using the enhanced So and Chan method, the steps of which include: step SC1: providing electrocardiogram data; step SC2: calculating a slope; step SC3: obtaining a slope maximum value; step SC4 : calculating a slope threshold; step SC5: confirming whether the slope is positive; if yes, proceeding to the next step; if not, calculating the absolute value of the slope, and confirming whether the absolute value of the slope of two consecutive points is greater than The slope threshold; if not, return to the step SC2; if yes, proceed to a step SC7; step SC6: confirm whether the slope of the two consecutive points is greater than the slope threshold; if not, return to the step SC2 If yes, proceed to the next step; step SC7: detecting a QRS complex starting point; step SC8: retrieving an R wave maximum extremum or an R wave minimum extremum; step SC9: delay time; and step SC10: If the slope is positive, the slope threshold is updated and the process ends.

In this way, the R wave detection error can be effectively reduced, the R wave detection accuracy rate is improved from 94.61% to 99.16%, and the FPGA development board is used to verify the R wave detection.

Steps of the SA1~SA8‧‧‧So and Chan algorithm

Steps of the steps SB1~SB11‧‧‧Modified So and Chan algorithm

Steps SC1~SC10‧‧‧ steps of the invention

Figure 1 is a flow chart showing a conventional So and Chan algorithm.

Figure 2 is a flow chart showing a conventional Modified So and Chan algorithm.

Figure 3 is a flow chart showing the R-wave detection algorithm of the present invention using the enhanced So and Chan method.

Figure 4 is a graph showing a comparison of the algorithm of the present invention with a conventional algorithm.

Fig. 5 is a graph drawn from the data of Fig. 4.

Fig. 6 through Fig. 13 show the results obtained by extracting data from the MIT-BIH database and performing simulation analysis by the method of the present invention.

Figure 14 shows the actual measured data.

Fig. 15 is a waveform in which the peak value of the T wave is higher than the peak value of the R wave in the waveform using the data of Fig. 14.

Figure 3 is a flow chart showing the R-wave detection algorithm of the present invention using the enhanced So and Chan method.

Please refer to FIG. 3, the R wave detection algorithm of the present invention using the enhanced So and Chan method, the steps of which include: step SC1: providing electrocardiogram data; step SC2: calculating slope; step SC3: Obtaining a slope maximum value; step SC4: calculating a slope threshold (slope_threshold); step SC5: confirming whether the slope is a positive value; if yes, proceeding to the next step; if not, calculating an absolute value slope_n of the slope (step SC51), and Confirming whether the absolute value of the slope of two consecutive points is greater than the slope threshold; if not, returning to step SC2; if yes, proceeding to step SC7; Step SC6: confirm whether the slope of two consecutive points is greater than the slope threshold; if not, return to step SC2; if yes, proceed to the next step; step SC7: detect the QRS complex start point (QRS onset); step SC8 : Searching for the R wave extreme value is maximum or minimum; Step SC9: Hold time; and Step SC10: If the slope is positive, the slope threshold is updated and the flow is ended.

Wherein, in the above step SC2, the So and Chan algorithm calculates the slope of each point of the acquired time domain data (calculates the slope slope(n) of the signal X(n) of each point n), as in the formula (1) Shown, slope ( n )=-2 X ( n -2)- X ( n -1)+ X ( n +1)+2 X ( n +2) (1).

Furthermore, calculating the slope threshold (slope_threshold) is as in equation (2),

The threshold parameter (threshold_parameter) can be adjusted to 2, 4, 8, 16 according to the voltage amplitude of the electrocardiogram; the threshold value represents the ratio of the magnification of the maximum slope max i in the ith second signal, and the smaller the parameter, the pair of R waves It has higher sensitivity when detecting, but has poorer ability to resist noise. The larger the parameter, the lower the sensitivity to R wave, but the stronger the ability to resist noise. Here, 8 is taken as an example. When the slope is greater than the slope threshold, it represents the starting point of the QRS wave (QRS onset), and the maximum value can be found to locate the R wave. Each time the R wave is detected, maxi is updated once. Equation (3) and formula (4) are as follows: First _maxi= Height of R po int- height of QRS onset (4).

The filter parameter (filter_param) can also be set to 2, 4, 8, and 16. Taking 16 as an example, the initial value of maxi is preset to the maximum slope value in the first second signal.

The following diagram illustrates the results obtained by the MIT-BIH database and the simulation results obtained by the method of the present invention. The verified electrocardiogram data is taken from the MIT-BIH database of the Massachusetts Institute of Technology, which has been widely used in recent years. MIT-BIH is a database of arrhythmia. Each piece of data is 30 minutes of data, which can be obtained directly from the MIT website. MIT-BIH Data 102 is the number 102 in MIT-BIH. ,the following The expression is analogous.

Please refer to Fig. 6, which shows that the data of MIT-BIH Data 102 is detected by the algorithm of the present invention, and the reverse R wave can be accurately captured. Please refer to Figure 7 for the reverse R-wave detection of the MIT-BIH Data 106 data. Please refer to Fig. 8 for the data of MIT-BIH Data 111. The data shows the situation of the left bundle branch blockage (LBBB) for a period of time. The algorithm of the present invention can also detect the first R. The peak will not be detected as the second R wave when the second peak appears. Please refer to Figure 9, which shows the data of MIT-BIH Data 111. This data has been blocked by the left bundle branch in the previous period. After a while, the right bundle branch blockage (RBBB) occurred. It is also true that only one R wave can be caught. Please refer to FIG. 10, which shows the sudden change of the position in the data of MIT-BIH Data 101. In the algorithm of the present invention, since the detected R wave is detected by the slope, if the ECG signal is encountered, Noise can be detected in the same way as baseline drift. Please refer to Figure 11, which shows the data of MIT-BIH Data 121, which has regular baseline drift noise and can detect the correct R wave. Please refer to Figure 12 for the information of MIT-BIH Data 109. It can be detected in a series of signals even if the amplitude of the R wave suddenly changes suddenly. Please refer to Fig. 13 for the data of MIT-BIH Data 209, in which a continuous R wave amplitude is high or low, and the algorithm of the present invention can still accurately capture the R wave. Please refer to Figure 14 for the actual measured data, and Figure 15 uses the data of Figure 14 for the waveform where the T-wave peak is higher than the R-wave peak, but with the Enhanced So and Chan algorithm of the present invention. The law still captures the R wave accurately.

Therefore, when an R wave is detected, the QRS complex start point (QRS onset) is paused for a period of time, and the R wave detection is not used to detect the QRS complex start point (QRS onset) The maximum value of the peak after a period of time, but the first turning point, this way can effectively improve the occurrence of false detection of RSR' wave and T wave too high; and reduce Modified So and Chan has been in Calculating the operation amount of the square of the slope is to judge the negative value of the negative value of the slope. This method can accurately detect the reverse R wave; and when the reverse R wave is detected, the threshold is not used. Update; in addition, it can effectively reduce the R wave detection error, increase the R wave detection accuracy from 94.61% to 99.16%, and use the FPGA development board to complete the R wave detection verification (such as Figure 4 and Figure 5).

Other or further aspects of the invention when the foregoing is directed to various embodiments of the invention The embodiments can be devised without departing from the basic scope and the basic scope is defined by the following claims. Although the present invention has been explained in connection with the preferred embodiments, it is not intended to limit the invention. It should be noted that many other similar embodiments can be constructed in accordance with the teachings of the present invention, which are within the scope of the present invention.

Steps SC1~SC10‧‧‧ steps of the invention

Claims (1)

An R-wave detection algorithm using an enhanced So and Chan method, the steps comprising: step SC1: providing electrocardiogram data; step SC2: calculating time-domain data of the electrocardiogram data, signal X of each point n (n) Slope slope(n), slope ( n )=-2 X ( n -2)- X ( n -1)+ X ( n +1)+2 X ( n +2); step SC3: obtaining a slope Maximum value; step SC4: calculating a slope threshold slope_threshold, Where, threshold_parameter is a parameter of the threshold value, which can be adjusted according to the voltage amplitude of the electrocardiogram data, the threshold value representing a ratio of the magnification of the maximum slope max i in the ith second signal; step SC5: confirming whether the slope slope(n) is Positive value; if yes, proceed to the next step; if not, calculate the absolute value of the slope slope(n) and confirm whether the absolute value of the slope of two consecutive points (n, n+1) is greater than the slope threshold If not, return to step SC2; if yes, proceed to step SC7; step SC6: confirm whether the slope of two consecutive points (n, n+1) is greater than the slope threshold; if not, return to the Step SC2; if yes, proceed to the next step; step SC7: detecting a QRS complex starting point; step SC8: retrieving an R wave maximum extremum or an R wave minimum extremum; step SC9: delay time; SC10: If the slope is positive, the slope threshold is updated and the process ends.